Table II.
Characteristics of included studies
| Study | Type of study | 1st author country | Study site (Country) | Cancer studied | Type of lesions studied | AI/ML protocol | Name of the device/technology | No of patients | Male, n (%) | Female, n (%) | Mean Age (years) | No of lesions studied |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Kok et al13, 1996 | Prospective, observational | Netherlands | Netherlands | Cervical Cancer (Screening) | Cervical smear | ANN-based DS tool | PAPNET | 91,294 | 0 | 91,294 (100) | NA | 91,294 |
| Chang et al14, 1999 | Prospective, observational | Taiwan | Taiwan | Prostate cancer | Multiple parameters | Multifactorial DS system | PCES | 43 | 43 (100) | 0 | 67 | 43 |
| Nieminen et al15, 2002 | Randomized, Prospective, observational | Finland | Finland | Cervical Cancer (Screening) | Cervical smear | ANN-based DS tool | PAPNET | 108,686 | 0 | 108,686 (100) | 44±10.3 | 108,686 |
| de Veld et al16, 2004 | Prospective, observational | Netherlands | Netherlands | Cancer of Oral Cavity | Oral mucosal lesion | PCA; ANN | Autofluorescence spectroscopy | 155 | NA | NA | 57±1 | 176 |
| Dreiseitl et al9, 2009 | Prospective, observational | Austria | Austria | Skin cancer | PSL | ANN-based DS tool | MoleMax II instrument with added decision support system | 458 | NA | NA | NA | 3,021 |
| Lucidarme et al17, 2010 | Prospective, observational | France | Multiple* | Ovarian cancer | TVS image of ovary | Not specified | OVHS | 264 | 0 | 264 (100) | 57 (Median) | 375 |
| Fink et al18, 2017 | Prospective, observational | Germany | Germany | Skin cancer | PSL | Not specified | MelaFind device | 111 | 59 (53.2) | 52 (46.8) | 45±17.3 | 346 |
| Mori et al8, 2018 | Prospective, observational | Japan | Japan | Colorectal cancer | Colorectal Polyps | Machine learning, SVM | Real-time automatic polyp detection system | 325 | 235 (72.3) | 90 (27.7) | 67 (Median) | 466 |
| Walker et al19, 2019 | Prospective, observational | USA | Israel | Skin cancer | PSL | CNN, Deep learning | NA | 63 | 34 (54.0) | 29 (46.0) | 50.4±14.9 | 63 |
| Wang et al20, 2019 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal Polyps | Deep learning architecture | Real-time automatic polyp detection system | 1,058 | 512 (48.4) | 546 (51.6) | 49.9±13.8 | 767 |
| Su et al21, 2019 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal polyps | CNN, Deep learning | AQCS-aided colonoscopy | 623 | 307 (49.3) | 316 (50.7) | NA | 442 |
| Li et al22, 2019 | Prospective, observational | China | China | Lung cancer | Lung nodules | CNN, Deep learning | DL-CAD | 346 | 221 (63.9) | 125 (36.1) | 51.0±10.2 | 1916 |
| Hollon et al23, 2020 | Prospective, observational | USA | USA | Brain cancer | Intraoperative surgical specimen | CNN, Deep learning | NA | 278 | NA | NA | NA | 278 |
| Wang et al24, 2020 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal polyps | Deep learning | CADe colonoscopy system | 369 | 179 (48.5) | 190 (51.5) | NA | 811 |
| Repici et al25, 2020 | Randomized, Prospective, observational | Italy | Italy | Colorectal cancer | Colorectal polyps | CNN, Deep learning | CADe colonoscopy system | 685 | 337 (49.2) | 348 (50.8) | 61.3±10.2 | 493 |
| Gong et al26, 2020 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal polyps | CNN, Deep learning | ENDOANGEL- assisted colonoscopy | 704 | 345 (49.0) | 359 (51.0) | NA | 369 |
| Wang et al27, 2020 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal polyps | Deep learning | CADe colonoscopy system | 962 | 495 (51.5) | 467 (48.5) | NA | 809 |
| Liu et al28, 2020 | Randomized, Prospective, observational | China | China | Colorectal cancer | Colorectal polyps | CNN, Deep learning | CADe colonoscopy system | 1026 | 551 (53.7) | 475 (46.3) | NA | 734 |
*Patients were recruited in five countries: France, Sweden, Italy, Germany, and Israel. AI/ML: artificial intelligence/machine learning; ANN: artificial neural network; AQCS: automatic quality control system; CADe: computer-aided detection; CNN: convoluted neural network; DL-CAD: Deep-learning based computer-aided diagnosis; DS: decision support; OVHS: ovarian histoscanning; PCA: principal component analysis; PCES: prostate cancer expert system; PSL: pigmented skin lesion; SVM: support vector machine; TVS: transvaginal scan. ENDOANGEL is a proprietary name